TextWorld: A Learning Environment for Text-based Games

TL;DR

TextWorld is a Python-based environment for training and evaluating reinforcement learning agents on customizable, procedurally generated text-based games, facilitating research on generalization, transfer learning, and game difficulty control.

Contribution

It introduces a flexible framework for creating, managing, and benchmarking text-based games, addressing challenges like partial observability and sparse rewards.

Findings

Baseline agents evaluated on curated and generated games.

Framework enables controlled variation in game difficulty and language.

Supports study of generalization and transfer learning in text games.

Abstract

We introduce TextWorld, a sandbox learning environment for the training and evaluation of RL agents on text-based games. TextWorld is a Python library that handles interactive play-through of text games, as well as backend functions like state tracking and reward assignment. It comes with a curated list of games whose features and challenges we have analyzed. More significantly, it enables users to handcraft or automatically generate new games. Its generative mechanisms give precise control over the difficulty, scope, and language of constructed games, and can be used to relax challenges inherent to commercial text games like partial observability and sparse rewards. By generating sets of varied but similar games, TextWorld can also be used to study generalization and transfer learning. We cast text-based games in the Reinforcement Learning formalism, use our framework to develop a set of benchmark games, and evaluate several baseline agents on this set and the curated list.